On Wed, Feb 02, 2000 at 11:03:13PM +0100, Frank Mittelbach wrote: > James, > sorry to disappoint you but I would like to answer both suggestions with no. > here are the arguments: > > > FMI: > > > > It kind of feels wrong to me to open up that name space just to > > > > be able to support \"A as an instance name --- but convince me > > > > otherwise :-) > > > > No but it should support Ä. > > it should not. template names are abstract names of the language and I > like the language portable and this means 7bit visible ascii plus a > few extra chars, just like label names. I'm not at all in favour of > supporting inputenc 8 bit in such places. At the moment, as I understand it, you can use anything which you could use in a \csname. This means that if you are not inputenc, or you are using T1 font encoding, you can use `Ä', as it either is, or expands to ^^c4. If portability is desired, and surely it is, `Ä' should either always work, or always not. I would prefer it to work. By the way, I'm not suggesting that the standard templates written by the latex3 project should use 8-bit characters, just that the template mechanism should not rule them out (of if it does, it rules them out completely) > > Many such problems should go away if a font encoding such as T1 or > > LY1 is used. Perhaps T1 should be made the norm for latex2e*? > > now the NO is the following: NO 2e* will not do that switch. 2e* (and > this is why it is called 2e*) will be a package based extension of 2e. > in other words it will run completely on a 2e kernel . I understand this, what I mean is that ldsetup.sty could select a fontencoding. I can see some problems with this, but nothing terrible. If it is in your mind to change this eventually I think this adds strength to my case. > open to discuss those points but note that I will vanish for a few > days to work on output routines and front matter. Super! Good luck. James